Ambient particulate matter attenuates Sirtuin1 and augments SREBP1-PIR axis to induce human pulmonary fibroblast inflammation: molecular mechanism of microenvironment associated with COPD

SREBP1 acts as the activated transcription factor upon PM treatment in HPF

To determine the exact canonical pathways or networking induced by PM in the human respiratory system, we evaluated transcriptomics datasets by microarray chips in HPF (Human Pulmonary Fibroblast) under several PM groups (sham, low dose for 6 hours, low dose for 24 hours, high dose for 6 hours, high dose for 24 hours, low dose = 5 μM, high dose = 50 μM) (Figure 1A). Through Genespring software normalization, 732 genes were found to have > 1.5-fold differential expression in low for 6 hours, low for 24 hours, high for 6 hours, or high for 24 hours relative to sham in HPF (Supplementary Table 1). The top ranking of candidate genes is shown in Figure 1B. The normalized microarray data was analyzed using IPA to identify significant molecular that are activated in response to PM exposure in HPF. The results showed that the transcription factor SREBP1 was the No. 1 putative candidates to be activated in response to PM exposure, with the transcriptional activity of its downstream genes of P = 7.9E-06 in HPF. The top 7 transcription factors in HPF are shown in Figure 1C. There are 2 SREBP genes in mammals, SREBP1 and SREBP2, encoding 3 sterol regulatory element-binding proteins (SREBPs), SREBP-1a, -1c and -2 [18]. SREBP-1a acts as a potent activator of all SREBP-responsive genes, stimulating cholesterol, fatty acids, and triglycerides synthesis. SREBP-1c is the main regulator of fatty acid, whereas SREBP-2 is the main regulator of cholesterol [18]. We sought to study if SREBP1 acts as transcription factors in response to PM, and our quantitative RT-PCR analysis result showed that SREBP1 was consistent with microarray in HPF (Figure 2A). SREBP1 was significantly increased in a dose-dependent manner after PM for 6 hours and in a time-dependent manner after PM at low dose. We uploaded our microarray datasets into the SREBP1-related signature from IPA tool, and our analysis showed that C-X-C motif chemokine ligand 8 (CXCL8), interleukin 6 (IL-6), Pirin (PIR), Interleukin 6 Family Cytokine (LIF), and Interleukin 24 (IL-24) were the most significantly upregulated transcriptional targets of SREBP1 (Figure 2B).

Figure 1. PM induced several potential transcription factors in a dose-dependent manner in normal lung cells. (A) The flowchart represents the procedure of microarray chips established in human pulmonary fibroblasts. (B) Detailed heatmaps highlight the significantly up- and down-regulated genes after PM treatment, that were normalized gene expression from the microarray database analysis. (C) The ranking of candidate transcription factors by IPA database of microarray from PM treatment compared with the sham group in human pulmonary fibroblasts. The cut-off was a 1.5-fold change.

SREBP1 activated PIR in response to PM exposure

We further examined the levels of SREBP1 and its downstream targets, CXCL8, IL-6, PIR, LIF, and IL-24 in the PM treatment groups (sham, low for 6 hours, low for 24 hours, high for 6 hours, high for 24 hours) of HPF. Our quantitative RT-PCR analysis results first showed that IL-6 and PIR were consistent with microarray in HPF, whereas CXCL8, LIF, and IL-24 were not consistent (Figure 2C). Second, response of PIR, IL-24, and LIF to PM was in a time- and dose-dependent manner. Response of CXCL8 to PM was found to be in a dose-dependent manner and only in a time-dependent manner at low dose. Response of IL-6 to PM was only in a dose-dependent manner for 6 hours and only in a time-dependent manner at low dose. Therefore, we assessed the expression of PIR at the protein level in response to PM treatment.

Figure 2. Signatures regulated by PM in HPF models. (A) The mRNA level of SREBP1 after PM treatment. (B) The prediction of SREBP1-network derived from the common signature, by comparison of IPA database with microarray data from HPF cells with a 1.5-fold-change cut-off. The intensity of the node color indicates the degree of activating (orange) or inhibiting (blue) regulation following PM interaction. (C) Several mRNA levels of SREBP1 downstream factors, including PIR, LIF, IL-6, IL-24, and CXCL8, after PM treatment in HPF.

Our western blot results in the PM treatment groups (sham, low for 6 hours, low for 24 hours, high for 6 hours, high for 24 hours) of HPF showed that PIR were significantly overexpressed after exposure to PM. PIR was significantly increased in a dose-dependent manner for 6 hours and only in a time-dependent manner at low dose (Figure 3A). At present, SREBP1 is a well-established negative target of SIRT1 [17, 19]. We further hypothesize that PIR is negatively associated with SIRT1.

Figure 3. PM exposure is positively correlated with activation of the SREBP1-PIR signaling pathway through SIRT1 downregulation. (A) Western blot analysis of PIR after PM treatment in HPF. (B) PIR activity from cell culture supernatant in PM-induced models. (C) Levels of SREBP1, SIRT1, and PIR were measured from cell culture supernatant of PM-induced models. (D) mRNA levels of SREBP1 in PF429242-PM co-treated HPF models. (E) mRNA levels of PIR in PF429242-PM co-treated HPF models. (F) PIR activity from cell culture supernatant in PF429242-PM co-treated HPF models. (G) Western blot analysis of PIR and SIRT1 in TPh A-PM co-treated HPF models. (H) PIR activity from cell culture supernatant in TPh A-PM co-treated HPF models. (I) mRNA level of PIR in Ex527-PM co-treated HPF models. (J) Western blot analysis of SIRT1 and PIR in Ex527-PM co-treated HPF models. (K) Level of PIR was measured from cell culture supernatant in Ex527-PM co-treated HPF models.

Increased PIR level and activity in cell culture supernatant are significantly correlated with PM exposure

PIR has been identified to possess enzymatic activity [20]. Moreover, delocalization of PIR was identified in melanomas [20]. To address the role of delocalization of PIR, PIR levels and PIR activity in cell culture supernatant of PM-induced models were determined. Our analysis showed that PIR enzymatic activity was significantly overexpressed after high-dose PM exposure (Figure 3B). Furthermore, we examined levels of SREBP1, SIRT1, and PIR in cell culture supernatant. Our result showed that SREBP1 and PIR were significantly increased in a dose-dependent manner, while SIRT1 was significantly suppressed in a dose-dependent manner after PM exposure, which was consistent with our hypothesis that PM exposure can activate the SREBP1-PIR axis via SIRT1 suppression (Figure 3C).

PM exposure can activate the SREBP1-PIR axis, which could be suppressed by SIRT1

In order to validate that PM exposure can activate the SREBP1-PIR axis, which could be suppressed by SIRT1, we administrate the SIRT1 inhibitors Ex527 [21], SREBP1 inhibitors PF429242 [22], and PIR inhibitors TPhA [23] in PM-treated HPF models. First, we administered PF429242 (10 and 100 μM) to PM-treated HPF models, and mRNA level of SREBP1 was significantly decreased as expected (Figure 3D). The mRNA level of PIR was significantly decreased in a dose-dependent manner in response to PF429242 in PM-treated HPF models (Figure 3E). We further examined cell culture supernatant in PF429242-PM co-treated models, and our analysis showed that PIR activity was significantly inhibited by PF429242 in a dose-dependent manner (Figure 3F). These results support the hypothesis that the SREBP1-PIR axis was activated by PM. Second, we administered TPh A (50 and 100 μM) to PM-treated HPF models, and protein levels of PIR were significantly decreased as expected (Figure 3G). Strikingly, the expression of SIRT1 was correlated with the expression of PIR on the protein levels in TPh A-PM co-treated HPF models (Figure 3G). We further examined the cell culture supernatant in TPh A-PM co-treated models, and our analysis showed that PIR activity was significantly inhibited in a dose-dependent manner as expected (Figure 3H). Third, we administered Ex527 (100 μM) to PM-treated HPF models, and our quantitative RT-PCR analysis showed that PIR was significantly overexpressed (Figure 3I). Afterwards, we further performed western blot analysis. We found Ex527 (50 and 100 μM) suppressed the expression of SIRT1 as expected, while Ex527 increased the expression of PIR in a dose-dependent manner (Figure 3J). Our analysis further showed that PIR in cell culture supernatant was significantly increased in a time- and dose-dependent manner in Ex527-PM co-treated models, which was consistent with expression on PIR protein level (Figure 3J, 3K).

NLRP3 inflammasome may be activated by PM exposure via SREBP1-PIR axis

Previous studies have identified that the protein complex called inflammasomes, play a central role in the acute lung injury and chronic pulmonary diseases [8]. NLRP3 and IL-1β were considered as characteristic markers of inflammasomes [8]. We first carried out quantitative RT-PCR analysis to detect the expression of NLRP3 in the 5 PM treatment groups (sham, low for 6 hours, low for 24 hours, high for 6 hours, high for 24 hours) of HPF, and our results showed that NLRP3 was overexpressed in dose-dependent manner in response to PM and only in a time-dependent manner at low dose (Figure 4A). We further carried out western blot to detect the expression of IL-1β and NLRP3 in the 5 PM treatment groups of HPF. After normalized against β-actin and quantified via ImageJ, NLRP3 was overexpressed in a dose-dependent manner for both 6 and 24 hours, while IL-1β was overexpressed in a dose-dependent manner for 6 hours and in a time-dependent manner at low dose (Figure 4B). Previous studies indicate that SREBP1, directly deacetylated and inhibited by SIRT1, increases expression of inflammasomes [17]. Thus, we hypothesized that exposure to PM suggests induction of the SREBP1-NLRP3 inflammasome pathway, via SIRT1 inhibition. We further validated the mRNA expression of NLRP3 in Ex527-PM and PF429242-PM co-treated HPF models. By quantitative RT-PCR analysis, Ex527 significantly upregulated NLPR3, while PF429242 significantly inhibited NLRP3 in a dose-dependent manner (Figure 4C, 4D). In Tph A-PM co-treated HPF models, the expression of PIR and SREBP1 were both downregulated in a dose-dependent manner, while NLRP3 was upregulated in a dose-dependent manner (Figure 4E).

Figure 4. NLRP3 inflammasome was upregulated by PM exposure. (A) mRNA level of NLRP3 after PM in HPF. (B) Western blot analysis of NLRP3 and IL-1β after PM in HPF. (C) mRNA level of NLRP3 in Ex527-PM co-treated HPF models. (D) mRNA levels of NLRP3 in PF429242-PM co-treated HPF models. (E) Western blot analysis of PIR, SREBP1, and NLRP3 in TPh A-PM co-treated HPF models.

PIR acts as a biomarker for detecting early COPD

To determine the clinical association between PIR and PM, we verified several previously developed microarray databases to identify and compare the correlation between PIR and nonsmokers, smokers, early COPD, and COPD using the Gene Expression Omnibus (GEO) databases. By analyzing a microarray cohort collecting small airway epithelium from phenotypically normal smokers and non-smokers, relative to nonsmokers, PIR was found to be the most activated gene to have a significant correlation with smoking (Figure 5A-5B, P < 0.0001) (GSE4498, n = 22). An in vitro study, assessing human airway epithelial cells, identified that air pollution stimulates a time- and dose- dependent increase in mRNA level of IL-6 via nuclear factor kappa B (NF-κB) mediated ROS [24]. Moreover, a multicenter longitudinal study that assessed 955 myocardial infarction survivors from six European cities identified a strong correlation between fibrinogen beta chain (FGB) and IL-6 and worse outcome after exposure to air pollutions [25]. Herein, we studied whether smoke exposure could induce genetic expression of FGB and IL-6. Our analysis showed that FGB and IL-6 had no significant association with smoking (Figure 5B, P = 0.0582 and P = 0.1572, respectively) (GSE4498, n = 22). In another microarray database comprised of large airway epithelium from phenotypically normal smokers, non-smokers, early COPD, and COPD, PIR was found to have a strong correlation with smoking (Figure 5C, GSE5060, n = 38). We further studied whether severity of COPD was correlated with genetic expression of PIR, FGB, and IL-6. Notably, relative to COPD, PIR was found to be the most activated gene to have a significant correlation with early COPD (Figure 5C-5D, P < 0.0001), followed by FGB (Figure 5D, P = 0.0006), while IL-6 had no significant association (Figure 5D, P = 0.3742) (GSE5060, n = 38). In order to validate if PIR is far superior in detecting early COPD, we next identified PIR protein expression using our own clinical cohort. By intra-tracheal instillation of PM into lung using mice models, we collected tissue sections from lung to undergo pathologic exam and immunohistochemical examination. The immunohistochemical staining results showed that the PM-treated mice group had higher concentration of the PIR protein in the lung tissue than in the control mice group (Figure 5E).

Figure 5. Overexpression of PIR correlates with smoke and early COPD. (A) Detailed heatmap highlights the correlation between the mRNA expression level of PIR and smoke in the GSE4498 cohort (n = 22) using Oncomine online analysis tool. (B) Differential mRNA levels of PIR, FBG, and IL-6 between non-smoker and smoker in the GSE4498 cohort. (C) Detailed heatmaps highlight the correlation between the mRNA expression levels of PIR and non-smoker, smoker, COPD, and early COPD in the GSE5060 cohort (n = 38) using Oncomine online analysis tool. (D) Differential mRNA levels of PIR, FBG, and IL-6 in non-smoker, smoker, COPD, and early COPD in the GSE5060 cohort (n = 38) in the analysis by the Oncomine online analysis tool. (E) The immunohistochemical staining results showed PM-fed mice group had higher concentration of the PIR protein in the lung tissue.

These data suggested that high PIR protein expression was significantly correlated with PM exposure, which means PIR acts as a predictor of PM-induced cardiopulmonary disease. Together, we identified that PM exposure can activate the SREBP1-PIR/NLRP3 inflammasome signaling pathway, through SIRT1 downregulation (Figure 6A, 6B). The SIRT1-SREBP1-PIR/ NLRP3 inflammasome axis may present several potential therapeutic targets for PM-related adverse health consequence.

Figure 6. A model illustrates that PM activates the SREBP1-PIR/inflammasomes signaling axis through SIRT1 -mediated modulation. (A) A schematic model to describe the positive feedback loop and detailed mechanisms of the SIRT1-SREBP1-PIR axis and crosstalk with NF-κB signaling. (B) SIRT1 modulates the SREBP1-PIR/NLRP3 inflammasome axis in response to PM. SIRT1 functions as a protector from PM exposure. The SIRT1-SREBP1-PIR/NLRP3 inflammasome axis may present an attractive therapeutic target for PM-related adverse health events.

PM

We purchased urban dust from Urban dust (1649b, NIST® SRM®, Gaithersburg, MD, USA). The dried PM samples were diluted with sterilized phosphate-buVered saline solution (PBS, pH 7.2) and then frozen at −80 C prior to use.

PM treatment

After reviewing the literature [53], we chose 5 μM PM as low dose and 50 μM as high dose. HPF was incubated with PM 5 μM, 50 μM, or sham for 6 hours and 24 hours. We estimated the global RNA expression for extra 6 hours, which is the optimal time point at which mRNAs and genes are fully transcribed by PM treatments. After PM incubation for 6 or 24 hours, we extracted HPF RNA and performed Affymetrix u133 plus 2.0 cDNA microarrays to detect the global and pre-defined gene and mRNA probe changes at the specific time points.

Cell culture and reagents

Human pulmonary fibroblasts (HPF) were purchased from ScienCell research laboratories (Cat. #3300, ScienCell, Carlsbad CA, USA). They were grown in fibroblast medium (FM, Cat. #2301, ScienCell, Carlsbad CA, USA) and stored at 37 °C and 5% CO₂. The medium was supplemented with 2% fetal bovine serum (FBS; Cat. #0010, ScienCell, Carlsbad CA, USA), 1% growth supplement (FGS; Cat. #2352, ScienCell, Carlsbad CA, USA) and 1% penicillin/streptomycin solution (P/S; Cat. #0503, ScienCell, Carlsbad CA, USA).

Reagents and compounds

Reagents used for the in vitro culture system include Ex527 (Cat. #E7034, Sigma-Aldrich®, Deisenhofen, Germany), PF429242 (Cat. #3354, Tocris Bioscience, Bristol, UK), TPh A (Cat. #TPh A ab144376, Abcam®, Cambridge, MA, USA), and Human Pirin ELISA Kit (Cat. #MBS280347, MyBioSource, San Diego CA, USA).

RT-qPCR

Total RNA was extracted using TRIzol reagent (Invitrogen, Carlsbad, CA) according to the manufacturer’s instruction. The RNA was quantified by NanoDrop (Thermo Fisher Scientific, Waltham, USA). Reverse transcription PCR (RT-PCR) and nested RT-PCR amplification were performed with a SuperScript III kit (Invitrogen, Carlsbad, CA) according to the manufacturer’s instruction. The expression level of Glyceraldehyde 3-phosphate dehydrogenase (GAPDH) was predominantly used as an internal standard for RT-PCR. The specific primer sequences are as follows: SREBP1 forward is 5′–GCAAGGCCATCGACTACATT-3′ and reverse is 5’-GGTCAGTGTGTCCTCCACCT-3′, C-X-C motif chemokine ligand 8 (CXCL8) forward is 5′-ATGACTTCCAAGCTGGCCGTGGCT-3′ and reverse is 5′-TCTCAGCCCTCTTCAAAAACTTCTC-3′, interleukin 6 (IL-6) forward is 5′-GGTACATCCTCGACGGCATCT-3′ and reverse is 5′-GCCTCTTTGCTGCTTTCAC-3′, Nod-like receptor protein 3 (NLRP3) forward is 5′-CTTCTCTGATGAGGCCCAAG-3′ and reverse is 5′- GCAGCAAACTGGAAAGGAAG-3′, interleukin 24 (IL-24) forward is 5′-TGTGAAAGACACTATGCAAGCTC-3′ and reverse is 5′-GTGACACGATGAGAACAAAGTTG-3′, Interleukin 6 Family Cytokine (LIF) forward is 5′-CCAACGTGACGGACTTCCC-3′ and reverse is 5′-TACACGACTATGCGGTACAGC-3′, and PIR forward is 5′-GAGCAGTCGGAAGGGGTTG-3′ and reverse is 5′-TTAACTCGGGTCTGCCAATGC-3′.

Western-blot analysis

Western blot analysis was performed with primary anti-PIR antibody (GeneTex, Irvine CA, USA), anti-NLRP3 antibody (GeneTex, Irvine, CA, USA), anti-SREBP1 antibody (GeneTex, Irvine CA, USA), anti-IL-6 antibody (Abcam, Cambridge MA, USA), anti-IL-1β antibody (GeneTex, Irvine, CA, USA), anti-SIRT1 antibody (Proteintech®, Rosemont, USA), and anti-β-actin antibody (Sigma-Aldrich, St. Louis, MO, USA). The bands corresponding to western blot were quantified using ImageJ.

RNA microarray analysis

Total RNA was corresponding extracted and purified from HPF samples via RNeasy Mini kit (Qiagen®, Valencia CA, USA) as the initial material and then qualified by Agilent model 2100 Bioanalyzer (Agilent Technologies, Palo Alto CA, USA). All RNA samples were labeled with the GeneChip 3′IVT Express Kit (Affymetrix, Santa Clara CA, USA), hybridized, washed, and scanned to the arrays with GeneAtlas Hybridization, Wash, and Stain Kit (Affymetrix, Santa Clara CA, USA), as described by the manufacturer. Normalization genes were identically represented on the GeneChip Human Genome U133 Plus 2.0 Array (Affymetrix, Santa Clara CA, USA). Expression of genes was then generated after log2 transformation and normalization with GeneSpring software (Agilent Technologies, Palo Alto CA, USA). To uncover the significance of genes for further analysis, cutoff values for gene expression results greater than ± 1.5-fold-change were further overlaid onto Ingenuity Pathway Analysis (IPA; QIAGEN, Valencia, CA, USA) to estimate canonical pathways, downstream effects, and potential upstream regulators. in response to PM.

In vivo model

6-8-week-old nonobese diabetic/severe combined immunodeficient gamma (NOD.Cg-Prkdc^scid Il2rg^tm1Wjl/SzJ, NOD-SCID γ, NSG™, JAX®, Harbor ME, USA) male mice (20-25 g body weight) and B6.Cg-Lep^ob/J (B6 ob JAX®, Harbor ME, USA) were used. To establish in vivo PM model, PM particles were diluted with 10 μl of PBS and then intratracheally administered into the pulmonary regions with Hamilton syringe. Body weight and behavior were monitored weekly.

Immunohistochemistry staining analysis

Briefly, tissue sections (5 μm) were dewaxed and rehydrated. The slides were incubated in 10 mmol/L citric buffer (pH 6.0). Antigen retrieval was done in Tris-EDTA buffer (pH 9.0) and microwaved for 15 minutes. Endogenous peroxidase activity was blocked by placing the sections in a methanol bath containing 0.5% H₂O₂. After blocking as above, the slides were incubated with primary antibody against target proteins, followed by biotin-conjugated secondary antibody polymer-HRP reagent The endogenous peroxidase activity was then visualized with diaminobenzidine tetrahydroxychloride (DAB) solution.

Public database

Analysis of Gene Expression Omnibus (GEO), a public cohort microarray database, was used to validate the correlation between expression of biomarkers and PM. The differentially expressed genes in the dataset were determined by nonparametric and t-test analysis.

Statistical analysis

Statistics analysis was performed with SPSS 17.0 software for Windows (SPSS, Chicago, Illinois, USA). P < 0.05 on a two-sided test was considered significant.

Ethics approval

Animal care and experiments were performed in compliance with 3R principles and approved by the Academia Sinica Institutional Animal Care and Utilization Committee (IACUC).